download_and_normalise()

MAYAFLUX_API std::span< const float > MayaFlux::Buffers::download_and_normalise	(	const std::shared_ptr< Core::VKImage > &	image,
		std::vector< uint8_t > &	raw_staging,
		std::vector< float > &	work,
		const std::shared_ptr< VKBuffer > &	gpu_staging
	)

Download a VKImage to CPU and return a normalised float span.

Downloads image into raw_staging via TextureLoom::download_data, then normalises into work via Kakshya::as_normalised_float.

raw_staging is resized to match the image byte footprint when the image dimensions or format change. work is the caller-owned scratch buffer reused across calls to avoid per-frame allocation.

Returns an empty span if image is null, not initialised, or its format has no ImageFormat mapping.

Parameters

image	GPU-resident source image.
raw_staging	Persistent byte buffer for the GPU download. Reuse across calls.
work	Persistent float buffer for normalisation output. Reuse across calls.
gpu_staging	Optional persistent staging buffer for device-local images. Reuse across calls.

Returns

Normalised float span pointing into work (or directly into the variant storage for float-format images).

Definition at line 454 of file StagingUtils.cpp.

{
    if (!image || !image->is_initialized())
        return {};
 
    auto& loom = Portal::Graphics::TextureLoom::instance();
 
    const auto fmt_opt = loom.from_vulkan_format(image->get_format());
    if (!fmt_opt)
        return {};
 
    const size_t byte_size = static_cast<size_t>(image->get_width())
        * image->get_height()
        * loom.get_bytes_per_pixel(*fmt_opt);
 
    if (byte_size == 0)
        return {};
 
    raw_staging.resize(byte_size);
    loom.download_data(image, raw_staging.data(), byte_size, gpu_staging);
 
    Kakshya::DataVariant variant { raw_staging };
    return Kakshya::as_normalised_float(variant, work);
}

References MayaFlux::Kakshya::as_normalised_float(), image, and MayaFlux::Portal::Graphics::TextureLoom::instance().

Referenced by MayaFlux::Buffers::ImageCVProcessor< T >::processing_function().

Here is the call graph for this function:

Here is the caller graph for this function: